Price dispersion and demand uncertainty: Evidence from US scanner data

I use the Prescott (1975) hotels model to explain variations in price dispersion across goods sold by supermarkets in Chicago. I extend the theory to accounts for the monopoly power of chains and for non-shoppers. The main empirical finding is that the effect of demand uncertainty on price dispersion is highly significant and quantitatively important: More than 50% of the cross sectional standard deviation of log prices is due to demand uncertainty. I also find that price dispersion measures are negatively correlated with the average price but are not negatively correlated with the revenues from selling the good (across stores and weeks) and with the number of stores that sell the good.